Quick rebuttal

[Great Bustard]

Just wanted to answer this point in a filled thread:

http://forums.dpreview.com/forums/read.asp?forum=1032&message=41527329

What's clear is that, all else equal, 36 MP will never look worse than 22 MP, and will usually look better. How much better depends on many variables.

No true. You fail to answer this question - how a blurry 36mp photo looks and prints better than a tack sharp 22mp photo?

Sorry, qianp2k -- all else equal (same lens, same aperture, same sensor size, same relative AA filter strength), more pixels will always resolve more detail than fewer pixels.

For example, a 36 MP sensor will resolve between 0 and 28% more linear detail over a 12 MP sensor -- it will never resolve less (all else equal, of course).

 

[studio311]

I admire your patience and perseverence Great Bustard.

I fear some will not be deterred from believing 22MP is the holy grail...a resolution of magical qualities...and any more, especially and specifically 36MP is just horrible.

Of course, that will all change once Canon releases a 40+mp camera next year.

 

[Silverback1988]

Sorry, qianp2k -- all else equal (same lens, same aperture, same sensor size, same relative AA filter strength), more pixels will always resolve more detail than fewer pixels.

For example, a 36 MP sensor will resolve between 0 and 28% more linear detail over a 12 MP sensor -- it will never resolve less (all else equal, of course).

Joe,
if we keep the noise constant (with applying NR),
then the resolution will be decrease as we raise the ISO.
Is that correct ?

For example,
say we get 36 MP at ISO 100,
and at ISO 1600 we could just get about 10 MP resolution ?

And I think the "resolution decrease" is not linear between ISO 100 to 400.
-
Brian

 

[Dylthedog]

I admire your patience and perseverence Great Bustard.

I fear some will not be deterred from believing 22MP is the holy grail...a resolution of magical qualities...and any more, especially and specifically 36MP is just horrible.

Of course, that will all change once Canon releases a 40+mp camera next year.

You are correct of course, just like when 12MP was the holy grail on a different forum until a 24MP machine dropped

 

[Rick Knepper]

My jab at it: Given the same surface area e.g. a FF sensor, more pixels per line (horizontal and/or vertical) records more/finer detail at capture due to a finer "resolution" (the non-print big definition of the word).

All any Canon user had to do was compare their 5D to the 5D2 before selling off their 5D, and this silly ongoing debate would have been virtually silent on this forum except from Nikon users (until now). Apparently few if any ever did. I did and the result was clear, more megapixels, more detail at my target image presentation size (1800x1200 electronically as viewed from 12 to 18 inches - essentially my computer monitor an NEC 2690). This is one part of the IQ equation but certainly not the only. For landscape styles, it is an important consideration, for other styles, other factors may be way more important.

I know that after comparing 12 MP to 21 MP, I saw an increase in captured detail. My only concern for the 21 MP to 36 MP jump is the law of diminshing return for my aforementioned target presentation size.

Of course, the possibility exists that these "arguments" are being made by folks who really know better but are currently under the control of fanboi compulsions.

This gentleman talking about blurry this and that must be referring to images shot by someone with poor technique (which has nothing to do with the camera). Also, he doesn't seem to understand the phrase "all else being equal".

http://forums.dpreview.com/forums/read.asp?forum=1032&message=41527329

What's clear is that, all else equal, 36 MP will never look worse than 22 MP, and will usually look better. How much better depends on many variables.

No true. You fail to answer this question - how a blurry 36mp photo looks and prints better than a tack sharp 22mp photo?

Sorry, qianp2k -- all else equal (same lens, same aperture, same sensor size, same relative AA filter strength), more pixels will always resolve more detail than fewer pixels.

For example, a 36 MP sensor will resolve between 0 and 28% more linear detail over a 12 MP sensor -- it will never resolve less (all else equal, of course).

--

Rick Knepper, photographer, photography never for sale, check my profile for gear list and philosophy.

 

[Rick Knepper]

I agree with you regarding smaller apertures in real world shooting but another characteristic of comparison testing you forgot to take into account is the use of "flat wall" scenes. There is no depth to the book shelf scene to make a difference shooting @ f11. Removing diffraction with larger apertures reveals more resolution but I suspect your intent here is to show that the 5D3 stands up better than the D800 at smaller apertuires due to less diffraction based on MPs.

davexl wrote:

100% Crops - but both uprezzed to 10,000 pixels, so that the 5D-II is not disadvantaged by being the only one resampled.

You're one of the few people that actually recognizes the fallacy of uprezzing one image to match the size of another image at its native resolution. Kudos for doing a proper comparison of images with different resolutions.

Anyway, almost all of the 5D3-D800 comparisons I've seen are based off tests at f8.0, which is of course the pixel-peeper's way of avoiding diffraction. Realistically, though, the f11 aperture is simply more practical in photos that would actually benefit from a high resolution sensor.

Whether it's shooting architecture or landscape, depth of field is a far bigger concern than pixel-level sharpness. On my 5D2, I never hesitate to shoot f16 whenever the situation calls for it. I look at all my landscape and architectural photos, and I seldom shot smaller than f11 (unless I was hand-holding shots at high ISO).

If I had a D800, I certainly won't be restricting myself to f8.0 only. Did you by any chance do a similar comparison at f11? I'd be curious to see how much of a difference there is between 36 MP and 22 MP.

--

Rick Knepper, photographer, photography never for sale, check my profile for gear list and philosophy.

 

[oysso]

http://forums.dpreview.com/forums/read.asp?forum=1032&message=41527329

What's clear is that, all else equal, 36 MP will never look worse than 22 MP, and will usually look better. How much better depends on many variables.

No true. You fail to answer this question - how a blurry 36mp photo looks and prints better than a tack sharp 22mp photo?

Sorry, qianp2k -- all else equal (same lens, same aperture, same sensor size, same relative AA filter strength), more pixels will always resolve more detail than fewer pixels.

For example, a 36 MP sensor will resolve between 0 and 28% more linear detail over a 12 MP sensor -- it will never resolve less (all else equal, of course).

It will not resolve less, but you loose FPS which for many are not a good thing. And you gain filesize... which is not always a good thing....

And many never print very large or crop very much, so 22 MP simply is enough for many...... Better with smaller files that is enough, rather than almost twice the size files which most information is never used anyway.

 

[qianp2k]

I cannot believe you start another thread just for that.

You still fail to answer this question - how a blurry 36mp photo looks and prints better than a tack sharp 22mp photo?

I don't think you understand what I said. I don't dispute D800 has more resolution than 5D3, of course. However as DPR said you need faster shutter in order to fully leverage 36mp resolution. For example under a non-ideal light, I can get a sharp 5D3 photo at ISO 200 at 1/20 hand-held while you likely must shoot under ISO 360 at 1/30 (just for an example may not exact number) in order to get similar sharpness when both viewed at 100% cropped. In another words, you need a better technique in order to achieve potential 36mp resolution. If you also shoot D800 in that scenario with ISO 200 at 1/20 you have higher chance to get a blurry photo.

Now assuming we both shoot ISO 200 at 1/20, I have sharp 5D3 photo while you have blurry D800 photo when both viewed at 100% cropped. Can you still say your blurry D800 photo is better than my sharp 5D3 photo, by either downsampling to 22mp or upsampling to 36mp? That's exactly what DPR means,

Can the D800 make good on its pixel count and provide a level of fine detail that trumps its DSLR rivals? It can. We emphasize the word can, because if you're truly after 36MP performance, be prepared to do some work. Flawless technique, fast shutter speeds and top-shelf equipment (particularly lenses and a tripod) along with a low ISO are requirements not options.

You need even more technique to leverage 80mp Mamiya MF camera.

So what this better technique for high pixel camera means? It means in order to have the best possible quality when print big such as 60" from D800, you really need to shoot under faster shutter or on tripod with the nice lens, at the base ISO and optimized F number (such as 5.6). Otherwise you probably waste D800 potential if you only print to 30x20". How many landscape photog using 40-80mp MP camera will shoot in hand-held?

http://forums.dpreview.com/forums/read.asp?forum=1032&message=41527329

What's clear is that, all else equal, 36 MP will never look worse than 22 MP, and will usually look better. How much better depends on many variables.

No true. You fail to answer this question - how a blurry 36mp photo looks and prints better than a tack sharp 22mp photo?

Sorry, qianp2k -- all else equal (same lens, same aperture, same sensor size, same relative AA filter strength), more pixels will always resolve more detail than fewer pixels.

For example, a 36 MP sensor will resolve between 0 and 28% more linear detail over a 12 MP sensor -- it will never resolve less (all else equal, of course).

--
http://qianp2k.zenfolio.com/

 

[orlandoom]

The following comes straight out of Nikon Technical Manual for the D800/E.

Introduction, Paragraph 2:
"While its high pixel count of 36 megapixels gives the D800/
D800E resolution unrivalled by previous digital SLR cameras,
a side effect is that bokeh and blur are made that much more
obvious. Realizing the full potential of a camera with over
30 million pixels involves a thorough appreciation of bokeh
and blur, careful selection of settings and of tools (such as
lenses and tripods), and working with the best possible subjects. "

Link to Manual

http://www.nikonusa.com/en_US/o/Y6wrkA9OU_z04IreazIXl_22UII/PDF/D800_TechnicalGuide_En.pdf

Don

 

[Ken Phillips]

But that 800E really has my number.

Canon ISO's seem to be a bit optimistic(comparing Mk. III to D800): you can shoot a faster shutter speed with the Nikon at a given aperture and ISO. That's a fact that will help a little bit with hand-held shots.

Also, diffraction and motion blur have no additional effect on a given print size due to the increase in pixel count in the 800 versus the Mk. III.

I have many Canon bodies and lenses, but I'll pick up an 800E and a couple of Nikon prime lenses (which, together, will be less than a spare 1Ds III), and use good technique to make best use of that resolution. I'd really like a monochrome version of that beast!
KP

 

[Macgupta]

This is because the increased pixels are increased, improved pixels. They have less noise than would be expected by a naive scaling down of the larger pixel to a smaller size. Without this advancement, the higher megapixel image would look like cr*p.

 

[Techblast]

Ken, thats what I was going to do. I love the 800E, but I don't love the Nikon Camera Body, poor LV implementation, focus calibration issues, Mirror Slap and vibration, lockups.

I'll just wait until these engineering issues are resolved by Nikon, and hopefully, by then Canon will have a better Hi-Res camera to offer, but we'll see. My 5DMKII is just fine for now.

 

[Faintandfuzzy]

We've always seen this going to higher MP cameras. I noticed is when moving from the 1Ds to the 1Ds2, and from the 40D to the 7D.

And by now, I'd hope realize that Qianp2k knowledge base is extremely limited. I'm surprised you even took the time to reply to him.

 

[qianp2k]

don't mention my name. idiot as you even don't understand what dispute is about.
--
http://qianp2k.zenfolio.com/

 

[gdanmitchell]

My jab at it: Given the same surface area e.g. a FF sensor, more pixels per line (horizontal and/or vertical) records more/finer detail at capture due to a finer "resolution" (the non-print big definition of the word).

All any Canon user had to do was compare their 5D to the 5D2 before selling off their 5D

I kept my 5D when I purchased a 5D2, and I print photographs from both on an in-house Epson 7900, so I think I have a pretty valid basis for comparing the two in the way you describe.

The higher MP camera does, indeed, have the capability to record finer resolution picture data the the lower MP camera. In very carefully shot photographs that taken through a very careful post processing workflow and then printed very large there are differences that may be visible to a person who knows what to look for and who looks very close. When printing on 13" x 19" paper, very few people would be able to consistently see a difference. Somewhere around 16" x 24" or so, the differences may start to be more visible to those who look carefully, and I feel much less confident about making a 24" x 36" print from a 5D than from a 5D2.

Do you print at 24" x 36" much? Do you always work very carefully from the tripod? If so higher MP is one of several factors that can make a difference. All else being equal, you can make a print a few inches larger with a 36MP original with the same pixel resolution, so if you shoot in the ways I described you might see the small differences I mentioned if you look carefully at your well-made 30" x 40" prints - assuming impeccable shooting technique.

Print at 30" x 40" very often?

Dan

--
---
G Dan Mitchell - SF Bay Area, California, USA
Blog & Gallery: http://www.gdanmitchell.com/
Google Plus: https://plus.google.com/u/0/102554407414282880001/
Facebook: http://www.facebook.com/gdanmitchellphotography
Flickr: http://www.flickr.com/photos/gdanmitchell/
Twitter: http://twitter.com/gdanmitchell

 

[Silverback1988]

has talent to use their gear to its maximum capabilities.
And I am sure that Joe is one of those people.
In other word,
I believe he knows how to make 36 MP full resolution from a 36 MP camera.

I think you make some valid points too in your post.
a blurry photo is more noticeable in high resolution.

And it is not easy to get a high resolution picture even if a high resolution gear is in your hand.

BUT Joe has valid point too.
What he said is true.
I believe you already knew that.

-
Brian

You still fail to answer this question - how a blurry 36mp photo looks and prints better than a tack sharp 22mp photo?

I don't think you understand what I said. I don't dispute D800 has more resolution than 5D3, of course. However as DPR said you need faster shutter in order to fully leverage 36mp resolution. For example under a non-ideal light, I can get a sharp 5D3 photo at ISO 200 at 1/20 hand-held while you likely must shoot under ISO 360 at 1/30 (just for an example may not exact number) in order to get similar sharpness when both viewed at 100% cropped. In another words, you need a better technique in order to achieve potential 36mp resolution. If you also shoot D800 in that scenario with ISO 200 at 1/20 you have higher chance to get a blurry photo.

Now assuming we both shoot ISO 200 at 1/20, I have sharp 5D3 photo while you have blurry D800 photo when both viewed at 100% cropped. Can you still say your blurry D800 photo is better than my sharp 5D3 photo, by either downsampling to 22mp or upsampling to 36mp? That's exactly what DPR means,

Can the D800 make good on its pixel count and provide a level of fine detail that trumps its DSLR rivals? It can. We emphasize the word can, because if you're truly after 36MP performance, be prepared to do some work. Flawless technique, fast shutter speeds and top-shelf equipment (particularly lenses and a tripod) along with a low ISO are requirements not options.

You need even more technique to leverage 80mp Mamiya MF camera.

So what this better technique for high pixel camera means? It means in order to have the best possible quality when print big such as 60" from D800, you really need to shoot under faster shutter or on tripod with the nice lens, at the base ISO and optimized F number (such as 5.6). Otherwise you probably waste D800 potential if you only print to 30x20". How many landscape photog using 40-80mp MP camera will shoot in hand-held?

http://forums.dpreview.com/forums/read.asp?forum=1032&message=41527329

What's clear is that, all else equal, 36 MP will never look worse than 22 MP, and will usually look better. How much better depends on many variables.

No true. You fail to answer this question - how a blurry 36mp photo looks and prints better than a tack sharp 22mp photo?

Sorry, qianp2k -- all else equal (same lens, same aperture, same sensor size, same relative AA filter strength), more pixels will always resolve more detail than fewer pixels.

For example, a 36 MP sensor will resolve between 0 and 28% more linear detail over a 12 MP sensor -- it will never resolve less (all else equal, of course).

--
http://qianp2k.zenfolio.com/

 

[Silverback1988]

Print at 30" x 40" very often?

Dan

Not exactly that ratio, but close to that area
Some of these :
60 cm x 90 cm ---> about more than 20 frames last year.
60 cm x 150 cm ---> about 10 frames or so
100 cm x 150 cm ---> less than 10 frames

I use 5DII & 7D.
I wait for a high res camera from Canon

-
Brian

 

[qianp2k]

I don't underestimate anyone. You know I am not disputing whether D800 has more resolution than 5D3/5D2, not at all, that is a clear fact. I only dispute you have to shoot more carefully and use better technique, use best lenses and tripod to leverage the potential 36mp resolution, in order to look really good when you print big such as beyond 40" wide.

Treat D800 as a MF camera as it already in MF territory. How many serious MF photog if any shoot hand-held? You know the best time to shoot landscape is not under bright noon sunlight but on much dimmer dawn and dusk light so tripod is almost necessary for serious photog to fully leverage 36mp D800 resolution. Sure you can shoot in hand-held and still look nice when print below 30x20" but then you waste D800 capability. My argument is that if you're not a serious photog and willing to lug best equipment around and shoot mostly in hand-held, better to stay with 5D3/5D2 or wait rumored 24mp D600.

I think you make some valid points too in your post.
a blurry photo is more noticeable in high resolution.

And it is not easy to get a high resolution picture even if a high resolution gear is in your hand.

BUT Joe has valid point too.
What he said is true.
I believe you already knew that.

-
Brian

You still fail to answer this question - how a blurry 36mp photo looks and prints better than a tack sharp 22mp photo?

I don't think you understand what I said. I don't dispute D800 has more resolution than 5D3, of course. However as DPR said you need faster shutter in order to fully leverage 36mp resolution. For example under a non-ideal light, I can get a sharp 5D3 photo at ISO 200 at 1/20 hand-held while you likely must shoot under ISO 360 at 1/30 (just for an example may not exact number) in order to get similar sharpness when both viewed at 100% cropped. In another words, you need a better technique in order to achieve potential 36mp resolution. If you also shoot D800 in that scenario with ISO 200 at 1/20 you have higher chance to get a blurry photo.

Now assuming we both shoot ISO 200 at 1/20, I have sharp 5D3 photo while you have blurry D800 photo when both viewed at 100% cropped. Can you still say your blurry D800 photo is better than my sharp 5D3 photo, by either downsampling to 22mp or upsampling to 36mp? That's exactly what DPR means,

Can the D800 make good on its pixel count and provide a level of fine detail that trumps its DSLR rivals? It can. We emphasize the word can, because if you're truly after 36MP performance, be prepared to do some work. Flawless technique, fast shutter speeds and top-shelf equipment (particularly lenses and a tripod) along with a low ISO are requirements not options.

You need even more technique to leverage 80mp Mamiya MF camera.

So what this better technique for high pixel camera means? It means in order to have the best possible quality when print big such as 60" from D800, you really need to shoot under faster shutter or on tripod with the nice lens, at the base ISO and optimized F number (such as 5.6). Otherwise you probably waste D800 potential if you only print to 30x20". How many landscape photog using 40-80mp MP camera will shoot in hand-held?

http://forums.dpreview.com/forums/read.asp?forum=1032&message=41527329

What's clear is that, all else equal, 36 MP will never look worse than 22 MP, and will usually look better. How much better depends on many variables.

No true. You fail to answer this question - how a blurry 36mp photo looks and prints better than a tack sharp 22mp photo?

Sorry, qianp2k -- all else equal (same lens, same aperture, same sensor size, same relative AA filter strength), more pixels will always resolve more detail than fewer pixels.

For example, a 36 MP sensor will resolve between 0 and 28% more linear detail over a 12 MP sensor -- it will never resolve less (all else equal, of course).

--
http://qianp2k.zenfolio.com/

--
http://qianp2k.zenfolio.com/

 

[magic786]

I have not posted here for a long time but I feel compelled to reply to this. I have both systems including a D3 with 24-70 lens which I love. I have to laugh at posts like this which quotes technical data as the holy grail. Sure, it can help in the overall impression in evaluating a camera. However, nothing in my mind beats trying out the camera yourself. I have been in a serious dilemma, just like many others here, as to which camera I should upgrade and I must say I got put off sometimes by all of the negative comments about the 5d3 and the hordes of d800 fans who are more interested in telling the world how good the d800 is in comparison to the competition.

I went out and tried the 5d3 and I must say I am immensely pleased at just how good it is. I skipped the 5d2 because I was not impressed with the white balance and from a number of reviews this is definitely an issue with the d800 which made me try the canon first and I have always preferred canon colours. I am also impressed at the focus accuracy as this where canon has finally caught up with nikon in this class. The exposures are fantastic and the high iso is incredible. This is not to take anything away from the d800 which I have not personally tried.
Now I am putting my fire suit on before someone tries to flame me!!

 

[gdanmitchell]

I'm with you on the obsessive focus on supposed technical data, absent much of any real-world notions of what the actual meaning of the data are. It probably isn't news to anyone who reads photography forums that some of impassioned discussions about this or that technical spec are roughly the equivalent of arguing about which car's blue paint is bluer than the others.

You sound like just the person who should be looking at the 5D3. While my experience shooting the 5D2 since shortly after it was released tells me that your concerns about its performance were perhaps a bit misguided, it is certainly true that the 5D3 improves on it in ways that will make it a worthy upgrade to the 5D for many shooters who can use its particular feature set.

As to the question of whether the Nikon D800 or the Canon 5D3 is the "better" camera, it is really an unimportant question to almost all photographers who might actually benefit from either of these tools. Both Nikon and Canon make excellent gear, and both the D800 and the 5D3 are really fine cameras. There are few situations in which the choice to use one over the other will make any significant difference to the quality of one's work. A Canon shooter who has a fleet of lenses and wants to upgrade is almost certain to find that a switch to Nikon in order own a D800 will not make much sense. And Nikon shooter with a fleet of Nikon lenses would gain little from switching to Canon. I'd hold to the position even if the feature sets of the two cameras were reversed.

Almost the only people who need to think about this much at all are those very rare photographers who do not already own a particular brand and who are already excellent photographers. Frankly, for almost everyone else, a coin flip to select brand would work just fine.

Dan

I have not posted here for a long time but I feel compelled to reply to this. I have both systems including a D3 with 24-70 lens which I love. I have to laugh at posts like this which quotes technical data as the holy grail. Sure, it can help in the overall impression in evaluating a camera. However, nothing in my mind beats trying out the camera yourself. I have been in a serious dilemma, just like many others here, as to which camera I should upgrade and I must say I got put off sometimes by all of the negative comments about the 5d3 and the hordes of d800 fans who are more interested in telling the world how good the d800 is in comparison to the competition.

I went out and tried the 5d3 and I must say I am immensely pleased at just how good it is. I skipped the 5d2 because I was not impressed with the white balance and from a number of reviews this is definitely an issue with the d800 which made me try the canon first and I have always preferred canon colours. I am also impressed at the focus accuracy as this where canon has finally caught up with nikon in this class. The exposures are fantastic and the high iso is incredible. This is not to take anything away from the d800 which I have not personally tried.
Now I am putting my fire suit on before someone tries to flame me!!

--
---
G Dan Mitchell - SF Bay Area, California, USA
Blog & Gallery: http://www.gdanmitchell.com/
Google Plus: https://plus.google.com/u/0/102554407414282880001/
Facebook: http://www.facebook.com/gdanmitchellphotography
Flickr: http://www.flickr.com/photos/gdanmitchell/
Twitter: http://twitter.com/gdanmitchell